The first Dutch SDHB founder deletion in paraganglioma – pheochromocytoma patients

Background Germline mutations of the tumor suppressor genes SDHB, SDHC and SDHD play a major role in hereditary paraganglioma and pheochromocytoma. These three genes encode subunits of succinate dehydrogenase (SDH), the mitochondrial tricarboxylic acid cycle enzyme and complex II component of the electron transport chain. The majority of variants of the SDH genes are missense and nonsense mutations. To date few large deletions of the SDH genes have been described. Methods We carried out gene deletion scanning using MLPA in 126 patients negative for point mutations in the SDH genes. We then proceeded to the molecular characterization of deletions, mapping breakpoints in each patient and used haplotype analysis to determine whether the deletions are due to a mutation hotspot or if a common haplotype indicated a single founder mutation. Results A novel deletion of exon 3 of the SDHB gene was identified in nine apparently unrelated Dutch patients. An identical 7905 bp deletion, c.201-4429_287-933del, was found in all patients, resulting in a frameshift and a predicted truncated protein, p.Cys68HisfsX21. Haplotype analysis demonstrated a common haplotype at the SDHB locus. Index patients presented with pheochromocytoma, extra-adrenal PGL and HN-PGL. A lack of family history was seen in seven of the nine cases. Conclusion The identical exon 3 deletions and common haplotype in nine patients indicates that this mutation is the first Dutch SDHB founder mutation. The predominantly non-familial presentation of these patients strongly suggests reduced penetrance. In this small series HN-PGL occurs as frequently as pheochromocytoma and extra-adrenal PGL.


Background
Paragangliomas occur as tumors of parasympathetically innervated head and neck paraganglia (HN-PGL), as intraabdominal and thoracic extra-adrenal paragangliomas of the sympathetic paraganglia, and as pheochromocytomas (PCC) of the adrenal medulla. Sympathetic paragangliomas may present clinically with hypertension, sweating and palpitations due to catecholamine excess, and especially in cases with extra-adrenal localization, they may be metastatic and aggressive. HN-PGL usually follows a mild course but may lead to significant morbidity due to compromised function of cranial nerves.
Although many paragangliomas are apparently sporadic (i.e. no known family history), many patients will carry a germline mutation, and worldwide up to 30% of all cases can be shown to have familial antecedents [1].
The identification in HN-PGL families of germline mutations of SDHD (succinate dehydrogenase, subunit D) [2] was soon followed by the identification of germline mutations in SDHB [3] and SDHC [4]. These three genes encode subunits of the mitochondrial tricarboxylic acid cycle enzyme, succinate dehydrogenase (SDH). SDH also acts as the complex II component of the electron transport chain, locating SDH at the center of cellular metabolism.
Despite the fact that SDH is thought to act solely as a unified protein complex, mutations of subunit genes lead to striking differences in clinical phenotype. While mutations of SDHD are associated predominantly with HN-PGL, frequently multifocal and generally non-metastatic,SDHB mutation carriers frequently present with PCCs and extra-adrenal paragangliomas, and mutations of SDHB are more often found in patients with aggressive, metastatic disease [5]. Until recently mutations of SDHC were exclusively associated with HN-PGL, but have now also been identified in patients with PCC [6,7].
The majority of mutations of the SDH genes described in the SDH mutation database [8]http://chromium.liacs.nl/ lovd_sdh/home.php are missense and nonsense mutations (n = 225). To date only ten distinct large deletions of the SDH genes have been described.
Although nearly all familial paraganglioma in the Netherlands is accounted for by the Dutch SDHD founder mutations p.Asp92Tyr and p.Leu139Pro [9], several Dutch families carrying an SDHB mutation were recently identified [10]. Here we describe the results of SDHB gene deletion scanning of 126 paraganglioma-PCC patients. Nine apparently unrelated Dutch patients all showed deletions of exon 3 of the SDHB gene. In order to determine if exon 3 is affected by a deletion hotspot, we proceeded to the molecular characterization of the deletion, mapping the breakpoint in each patient, and used haplotype analysis to determine whether the patients share any common haplotype, which would suggest a single novel founder mutation. In addition, the clinical phenotype of these patients is described.

Patients
Between 2000 and 2008, a total of 251 index patients with either a paraganglioma or PCC were referred for molecular testing of the SDHD/B/C genes to the Molecular Genetics Laboratory at the Leiden University Medical Center, The Netherlands. Informed consent was obtained for DNA testing according to protocols approved by LUMC Ethics Review Board. DNA was available from 126 index patients who tested negative for SDHD point mutations and in whom point mutations of SDHB and SDHC were, in most cases, also excluded.

Multiplex ligation dependent probe amplification
MLPA was carried out with the P226 MLPA kit http:// www.mrc-holland.com, containing probes for all exons of the SDHB, SDHC and SDHD genes, as well as probes located in the promoter of each gene (27 different probes). MLPA analysis was performed according to the MRC Holland protocol [11] except that all reagents in the kit were used at 1/2 of the recommended volume and hybridization time was reduced from 16 to 2.5 hours. No difference in results was seen compared to recommended conditions.

Haplotype analysis
Analysis of haplotypes by polymorphic di-and tetra nucleotide markers (microsatellite markers) was performed according to standard procedures (details available upon request), using the following markers: D1S436, D1S2697, D1S170, D1S3669, D1S2826 and D1S2644. The distance between the last and the first marker is ~3.35 Mb. In addition, intragenic SNPs were sequenced to refine the haplotype ( Table 1). The frequency of marker alleles in the Dutch population was determined in 24 healthy controls.

Breakpoint characterization
Long range PCR using the primers F2 (5'-TCT GTT GTG CCA GCA AAA TG-3') and R4 (5'-CAA ATC CTG CCC TGA AAA AC-3') was carried out using the Takara LA Taq kit (Takara Bio Inc., Lucron Bioproducts B.V., Gennep, The Netherlands) following the manufacturers recommendations. The resulting PCR fragment of 8.5 kb in the patients carrying the deletion was subjected to restriction mapping using the following twelve enzymes: Alw44I, BglI, BglII, BsaHI, BspEI, EcoRI, Mph1103I, NdeI, SacI, ScaI, SmaI, XbaI. Analysis of the resulting restriction patterns narrowed the specific region of the deletion, and was followed by the design of primers 2162 (5'-CCA GTC CAT GAA AGG CAA-3') and 2164 (5'-GCT CCA TGT GTC ACG TGT TT-3'). This allowed the amplification of a 1.6 kb fragment in patients carrying the deletion but not in healthy controls. The PCR product was sequenced, and analyzed using the Multalin program http://bio info.genopole-toulouse.prd.fr/multalin/multalin.html. Sequence analysis of the SDHB gene was performed according standard procedures (details available upon request), using the NT_004610.18 reference sequence.

MLPA analysis
MLPA allows the detection of large deletions, and is based on the quantification of multiplexed amplified DNA fragments. It was first described by Schouten et al. in 2002 [11] and due to reliability and ease of use has been widely adopted in both research and diagnostic settings.
All 251 index patients were screened for germline mutations in SDH genes by sequence analysis. A pathogenic mutation was identified in 125 patients (50%). MLPA analysis was performed in 126 mutation negative patients, and included all exons and the promoter of the SDHB, SDHC and SDHD genes. Deletions of the SDHB gene were detected in nine patients, all affecting exon 3 (figure 1). Four deletions were identified in the SDHC and SDHD genes and will be described elsewhere.  An example of MLPA analysis of the SDHB, SDHC and SDHD genes

Breakpoint characterization
Restriction mapping of the 9 kb SDHB fragment was carried out using twelve enzymes. Analysis of restriction patterns narrowed the breakpoints to small regions of intron 2, approximately 4.5 kb upstream of exon 3 and to a region 3.5 kb downstream, in intron 3. Primers were designed around the expected site of the deletion, and a 1.6 kb fragment could be amplified in all patients carrying the deletion, indicating that all deletions were either identical or mapped to a small and specific region, which would suggest a mutation hotspot.
Sequencing of the 1.6 kb PCR product revealed identical breakpoints in all samples, resulting in a deletion of 7905 bp, including exon 3 (figure 3), suggesting a single founder mutation, identical by descent. Following HGVS cDNA nomenclature, the deletion is correctly described as c.201-4429_287-933del. The deletion of exon 3 is predicted to result in a frameshift at the DNA level and a truncated protein, p.Cys68HisfsX21.
Although analysis of the surrounding sequence revealed that the upstream breakpoint is located in an AluSz repeat, the downstream breakpoint is located in unique sequence, rather than in an Alu repeat. Thus there are no repeat sequences or sequence similarity around the breakpoint that would suggest a mechanism of deletion or location of a mutational hotspot.

Haplotype analysis
Although the patients had no apparent family connection, all are natives of the Netherlands and share an identical SDHB deletion.
Therefore we carried out haplotype analysis with six diand tetra nucleotide polymorphic markers surrounding the SDHB gene. Haplotyping was refined by typing additional intragenic SNPs. A common haplotype could be deduced in all patients ( figure 4), indicating descent from a common ancestor. The marker haplotype formed by D1S436, D1S2697, and D1S170 has a frequency of 1% in the Dutch population, and the likelihood of nine unrelated cases carrying this haplotype is ~1.4 × 10 -18 . Certain haplotypes have apparently mutated or recombined in selected patients, an indication that they are only distantly related. Clinical details for all nine patients are shown in overview ( Table 2).

Discussion
While the majority of SDH-related hereditary HN-PGL and PCC patients carry missense and nonsense mutations, a significant proportion of patients may carry whole gene or exon deletions. In the entire current series of patients 13 of 251 (5%) were found to have deletions, representing 10% of all mutations found. These data have clear implications for diagnostic DNA screening and indicate that deletion scanning such as MLPA should be considered once patients have tested negative by sequencing. ing the first whole gene deletion of SDHB and a deletion of exon 1. This latter deletion was shown to be a founder on further characterization and an additional unique exon 1 deletion was identified in a French family [15]. These patients showed predominantly retroperitoneal PGL, but two cases from the French family developed adrenal PCC.
Amar et al. have recently described a novel deletion of SDHB in a patient studied in relation to metastatic paraganglioma [16]. This deletion was not further characterized. Fish [19].
The clinical features of the SDHB exon 3 deletion patients described in this report differ from most of the reports described above in the higher than expected frequency of HN-PGL. Previous studies have shown that SDHB mutations predominantly predispose to abdominal or thoracic paragangliomas and adrenal PCC [20], but as Leiden is a national referral centre for HN-PGL, a referral bias may be operating.
Although relatively few deletions have thus far been described, the current picture does not indicate that the phenotype caused by deletions differs substantially from that of missense and nonsense mutations. While mutations of SDHB and SDHD show an intriguing divergence in related phenotypes despite the intimate association of the protein subunits, gene deletions are unlikely to show specific genotype-phenotype effects, relative to truncating or missense mutations, because all mutations are assumed to result in loss of protein function.
The best current estimates of penetrance for SDHB mutations are 77% at 50 years [20] and approximately 60% at 50 yrs [5]. These figures are based largely on index cases, as few patients present in the context of extended families. Such estimates are known to exaggerate penetrance, and more accurate estimates will require large SDHB-related families, and the inclusion and detailed clinical screening Haplotype analysis of patients carrying the SDHB exon 3 deletion Figure 4 Haplotype analysis of patients carrying the SDHB exon 3 deletion. Microsatellites flanking SDHB and intragenic SNPs demonstrate a common haplotype. The common haplotype and the region deleted are indicated.
of apparently unaffected mutation carriers in addition to patients.
The only large family described to date [21] suggests a much lower penetrance but there was no detailed clinical screening of mutation carriers and the authors made no attempt to estimate penetrance. The fact that SDHB germline mutation carriers often present as apparently nonfamilial cases has not escaped various authors [5,[22][23][24] and even proven founder mutations of SDHB may initially present as isolated families [15] (and this study).
In contrast, SDHD-related mutations show a very high penetrance, over 80% at 50 years [5], and also show a striking and unique imprinted or parent-of-origin inheritance [25], showing almost complete penetrance with paternal inheritance, while mutation carriers via the maternal line remain tumor-free throughout life. This is in sharp contrast with the SDHB and SDHC genes, located on chromosome 1, which do not show parent-of-origin inheritance. All evidence indicates that it is not the SDHD gene itself which is imprinted [2,26] but that some additional locus on chromosome 11 is involved.
The major imprinted locus of the human genome is on chromosome 11p15.5, and we have proposed a model in which an imprinted and maternally expressed gene on chromosome 11p15.5 must be lost together with the nor-mal maternal SDHD allele (the paternal allele is inactivated by a germline mutation) prior to initiation of tumorigenesis [26]. This mechanism, recently referred to as the 'Hensen model' [27], is supported by additional data for chromosome 11 [28] and may also play a role in chromosome 3-linked VHL PCC [29,30].
This model could provide a genetic explanation for the reduced penetrance of SDHB mutations, in which both SDHB and the modifier must be lost, requiring loss on two separate chromosomes; intrinsically less likely than a single genetic event (whole chromosome loss) which has been shown to be the mechanism in SDHD tumors [26,28]. Alternatively or additionally, the difference in penetrance may have a (partly) biochemical explanation, related to the difference in function of the SDHD and SDHB subunits, the former principally structural, the latter catalytic. A unique biochemical effect of SDHB mutations seems likely in the light of the differing location and often aggressive behavior of tumors, and the high number of clinically penetrant mutations identified. No current model of PGL tumorigenesis can explain these phenomena [31,32].

Conclusion
We describe the first Dutch founder mutation of SDHB, a novel deletion of exon 3. Index patients presented with PCC, extra-adrenal PGL as well as HN-PGL. Lack of a clear family history in seven out of nine cases strongly indicates reduced penetrance. Family studies will be extended to further delineate penetrance and expression.